PgCache vs Redis and Valkey

Redis is the usual way to cache Postgres reads: the app checks Redis first, queries Postgres on a miss, and clears keys when the data changes. Valkey, the Linux Foundation fork of Redis, works the same way. PgCache caches the same reads at the database connection, with no cache code in the app.

See which of your own queries PgCache would serve: check your workload or open a sample report.

Redis and Valkey facts checked against their docs and license pages on October 15, 2026.

PgCache

A caching proxy that speaks the Postgres protocol. Your app keeps sending SQL. Repeated reads come back from cache, and logical replication keeps the cached data current.

No cache code in the app
Writes from any app, job or tool reach the cache
Each connection reads its own writes

Redis / Valkey

In-memory key-value stores. Your app decides what to store, reads it back, and clears or expires it when the data changes.

Where they differ

Topic	PgCache	Redis / Valkey
App changes	Point the connection string at PgCache	Cache code around each cached query and each write
What it caches	SQL query results, picked from live traffic	Whatever the app stores under a key
Freshness	Within typical logical replication lag	Stale until the app clears the key or its TTL expires
Writes from other services and tools	Reach the cache through replication	Missed unless each one also clears the right keys
Queries	SQL, including joins and aggregates	Key lookups

Which one fits

Choose PgCache when

The reads you want to cache are SQL queries against Postgres
Several services or jobs write to the same tables
Nobody wants to own cache invalidation code

Choose Redis or Valkey when

You cache things that aren’t query results, like rendered pages or API responses
You need sessions, rate limits, queues or pub/sub

Valkey, Memcached and ElastiCache

Valkey is a fork of Redis that the Linux Foundation started in 2024, after Redis changed its license. Memcached is another in-memory key-value cache. Your app caches Postgres reads with Valkey or Memcached the same way it does with Redis.

On AWS, Amazon ElastiCache runs Valkey, Redis OSS or Memcached as a managed service. See PgCache vs Amazon ElastiCache.

What they have in common

Both take repeated reads off Postgres
Neither speeds up writes

Using them together

Redis or Valkey for sessions, queues and anything that isn’t a query result
PgCache for the SQL reads you’d otherwise cache by hand

Common questions

Can PgCache replace Redis?

For caching Postgres query results, yes, with no cache code in the app. Sessions, queues and other data that doesn’t live in Postgres stay in Redis or Valkey.

Can I keep Redis and add PgCache?

Yes. PgCache sits between your app and Postgres and doesn’t touch Redis.

Can Redis be kept current from Postgres automatically?

Yes, if you build a pipeline for it. Debezium streams row changes out of Postgres, and code you write updates Redis. PgCache vs Debezium and Redis compares that setup.